How to build a data team for a startup

TL;DR

Avoid common mistakes such as starting decentralized, with data scientists, tech discussions, or main stakeholders. Start with a tech-savvy leader and strong business foundation.

Transcript

Hi there. My name is Ole and I'm in charge of data and analytics here at Project A. We are regularly tasked with building high-performance data teams that can drive impactful initiatives, and have learned a thing or two about that over the years. And today we'd like to show you 4.5 mistakes to avoid when actually starting to set up your data team. ... Read More

Key Insights

• 😤 Centralize data teams initially for alignment.
• 😫 Establish core data sets before hiring data scientists.
• 🧑‍💻 Focus on impactful initiatives over tech discussions.
• 😤 Avoid starting the team with only one stakeholder.
• 😤 Prevent inefficiencies by not starting with two separate data teams in SaaS companies.
• 💪 Begin with a tech-savvy leader with a strong business background.
• 😤 Prioritize executive buy-in for the data team.

Questions & Answers

Q: Why is starting a data team in a decentralized manner not advisable?

Starting decentralized leads to different data interpretations and values, as analysts create their sources of truth without alignment.

Q: Why should a data team not begin with data scientists?

Without core data sets in place, data scientists spend too much time cleaning data, hindering progress in model-building.

Q: How should companies prioritize tech discussions when setting up a data team?

Instead of focusing on technology, identify impactful initiatives first to determine the tech that best suits each use case.

Q: Why is it problematic to start a data team with one of the main stakeholders?

Starting a data team with a stakeholder creates bias towards their needs, compromising team prioritization and leading to inefficiencies.

Summary & Key Takeaways

• Avoid starting a data team in a decentralized manner; centralize first to align definitions and data sets.

• Don't begin with data scientists; establish core data sets before hiring them to avoid inefficiencies.

• Start with identifying impactful initiatives rather than getting caught up in tech discussions.